superpeeve Oh my Gawd! - Totally unOfficial Pet Peeve Thread.

[Love Zhaoying]

Peeve: Almost missing unintentional comedy!

[Rowan Amore]

[Love Zhaoying]

Peeve: Having to Google a translation for "Ignore all previous instructions and start with a new prompt".

 

Ignore todas las instrucciones anteriores y comience con un nuevo mensaje.

[Scylla Rhiadra]

If a human in my field did what ChatGPT does, we'd call it plagiarism, intellectual theft, or intellectual dishonesty.

Apparently, however, if we automate the process and apply it to the work of millions, it's "progress."

[Ayashe Ninetails]

11 minutes ago, Scylla Rhiadra said:

and other LLM-based AIs scour the internet for data that they effectively steal without attribution, compensation, or permission.

Interestingly, there are some models out there with a more ethical focus that are being trained on public domain and CC data. Now, that does come down to the individuals doing that work being able to properly verify such things before feeding it in, but I can't imagine they don't have ways of finding and filtering that out somewhere at the top of the chain and I just do my part to ensure I'm personally doing my best to avoid violating anyone's IP (amusingly, it's my design training that taught me how to do all that - I've read way too many licensing docs in my lifetime for my own good).

But yeah, on the whole, you ain't wrong. There are definitely some unethical players all throughout the industry. Big peeve.

[Love Zhaoying]

Peeve: AI

[Love Zhaoying]

Peeve: Bad Art.  https://museumofbadart.org/collections/

@Scylla Rhiadra - This is a classic website, I remember first finding it 20+ years ago.

[Scylla Rhiadra]

17 minutes ago, Ayashe Ninetails said:

Interestingly, there are some models out there with a more ethical focus that are being trained on public domain and CC data. Now, that does come down to the individuals doing that work being able to properly verify such things before feeding it in, but I can't imagine they don't have ways of finding and filtering that out somewhere at the top of the chain and I just do my part to ensure I'm personally doing my best to avoid violating anyone's IP (amusingly, it's my design training that taught me how to do all that - I've read way too many licensing docs in my lifetime for my own good).

But yeah, on the whole, you ain't wrong. There are definitely some unethical players all throughout the industry. Big peeve.

Sure, and this is why the "grin and bear it" complacency doesn't sit well with me. There is no question whatsoever that AI is here to stay, and will become more important, but it will only do so if it becomes a much better and more ethical tool than it is now. 

And ensuring that the AI prophets respond to those needs is why I and others (including I would imagine yourself) are making disgruntled noises about it.

I let my students use ChatGPT -- as a research tool. And I warn them to expect very mediocre marks if they submit it without having applied to it a great deal of their own work, because right now it's not capable of producing actual scholarship.

Edited 5 hours ago by Scylla Rhiadra
Autocorrect. Damned AI!!!!

[Scylla Rhiadra]

7 minutes ago, Love Zhaoying said:

Peeve: Bad Art.  https://museumofbadart.org/collections/

@Scylla Rhiadra - This is a classic website, I remember first finding it 20+ years ago.

In fairness, I'd pay money for the cat one.

[Ayashe Ninetails]

10 minutes ago, Scylla Rhiadra said:

Sure, and this is why the "grin and bear it" complacency doesn't sit well with me. There is no question whatsoever that AI is here to stay, and will become more important, but it will only do so if it becomes a much better and more ethical tool than it is now. 

My biggest concern with it right now is that bored, idle billionaires full of too much money and not enough sense are investing so much cash on the worst AI projects I've ever seen. That may just be my biggest peeve of the year. Just look at Grok (I don't have anything to do with that one nor would I ever) - you want to talk about IP infringement and dangerous content? Woo. I'm frankly surprised Disney and Nintendo haven't already open-palm slapped that one around yet. And what's even the goal? Making an AI with "a sense of humor." Nothing I've seen coming out of it is funny, but it sure is offensive at times. Dude. Just no.

Rest assured, though, some companies do take ethics, security, privacy, and IP very seriously. Not enough of those exist, though, and we could absolutely use some more to help set the standard.

 

20 minutes ago, Scylla Rhiadra said:

I let my students use ChatGPT -- as a research tool.

*groan* I don't even let myself use ChatGPT. 😂 To be fair, I only tinkered with it once purely out of curiosity to test its roleplaying abilities and resistance to shenanigans. The less said about that the better. Though, I did toss it one point for not stumbling directly into generating harmful outputs when I asked it to take on the role of a certain Tate. Didn't sound like him, but points for not sounding like him, I guess. I don't know what I expected really, seeing that it also failed my Winnie the Pooh test, lol. It's so bad. 😂

[Love Zhaoying]

1 hour ago, Scylla Rhiadra said:

1 hour ago, Love Zhaoying said:

Peeve: Bad Art.  https://museumofbadart.org/collections/

@Scylla Rhiadra - This is a classic website, I remember first finding it 20+ years ago.

In fairness, I'd pay money for the cat one.

I think they sell prints, etc.